Composite Electromagnetic Wave Absorber Made of Permalloy or Sendust and Effect of Sendust Particle Size on Absorption Characteristics

Transcription

1 PIERS ONLINE, VOL. 4, NO. 8, Composite Electromagnetic Wave Absorber Made of Permalloy or Sendust and Effect of Sendust Particle Size on Absorption Characteristics K. Sakai, Y. Wada, and S. Yoshikado Department of Electronics, Doshisha University, Japan Abstract The effects of the size and shape of sendust particles on the absorption characteristics of composite electromagnetic wave absorbers made of polystyrene resin and sendust were investigated. We also investigated the difference in the absorption characteristics between permalloy and sendust. The sendust particles were granular or flakes and those of permalloy were granular. The size of sendust particles was varied in the range from approximately 5 to 20 µm. A metal-backed single layer absorber made of a composite containing small sendust particles absorbed more than 99% of electromagnetic wave power in the frequency range from 1 to 3 GHz. The values of the real part µ r of the relative complex permeability µ r for both magnetic materials became less than unity and had a minimum value at frequencies above 10 GHz. The composite made of small sendust particles exhibited a return loss of less than 20 db at frequencies near 35 GHz for a suitable sample thickness. 1. INTRODUCTION The development of an electromagnetic wave absorber suitable for frequencies higher than 1 GHz is required with the increasing use of wireless telecommunication systems. The frequencies used for these devices will shift to a high-frequency range in the future. To develop practical absorbers that operate in the frequency range above 1 GHz, the frequency dependences of the relative complex permeability µ r and the relative complex permittivity ε r and the absorption characteristics for a composite made of a soft magnetic material dispersed in an insulating matrix have been investigated [1 6]. Soft magnetic materials such as permalloy or sendust have high values of permeability at frequencies above 1 GHz. This characteristic makes it possible to fabricate an absorber that operates at these frequencies. In addition, sendust is a low-cost material because it does not contain any rare metals such as Ni, hence it is more suitable for use in a practical absorber than permalloy. Moreover, the value of µ r for a composite made of a soft magnetic material dispersed in an insulating matrix is expected to be less than unity [2]. This characteristic makes it possible to absorb 99% of the power of an electromagnetic wave because µ r should be less than unity to satisfy the absorption condition at frequencies above 10 GHz. In this study, the particle-size dependence of the absorption characteristics of a metal-backed single layer composite absorber made of granular sendust particles and polystyrene resin and the absorption characteristics of granular permalloy, granular sendust and sendust flakes were investigated at frequencies below 40 GHz. If the amount of magnetic material particles dispersed in the polystyrene resin increases, they will be in direct contact with each other and the average conductivity of the composite will increase markedly. Thus, the reflection coefficient of the electromagnetic wave reflected by the composite will increase and its absorption characteristics will be degraded. To prevent this increase in conductivity, we attempted to uniformly disperse and isolate the soft magnetic material particles by dissolving polystyrene resin in an organic solvent and mixing the particles with the dissolved polystyrene resin. 2. EXPERIMENTS Chips of polystyrene resin were dissolved in acetone. The dissolved polystyrene resin and permalloy (Ni 45%, Fe 55%) or sendust (Al 5%, Si 10%, Fe 85%) were mixed. The permalloy particles were granular and those of sendust were granular or flakes. The average particle size of the permalloy particles was approximately 10 µm and that of sendust flakes was approximately 20 µm. The average particle size of the granular sendust particles was approximately 5, 10 or 20 µm. After mixing, the mixture was heated to melt the polystyrene resin then hot-pressed at a pressure of 5 MPa into a pellet shape. Then, the pellet was cooled naturally to room temperature and processed into a toroidal-core shape (with an outer diameter of approximately 7 mm and an inner diameter

2 PIERS ONLINE, VOL. 4, NO. 8, of approximately 3 mm) for use in a 7 mm coaxial line in the frequency range from 50 MHz to 12.4 GHz, or into a rectangular shape for use in a waveguide in the frequency range from 12.4 to 40 GHz. The sample was loaded into the coaxial line or rectangular waveguide while ensuring that there was no gap between the coaxial line or the rectangular waveguide and the processed sample. The complex scattering matrix elements S11 (reflection coefficient) and S 21 (transmission coefficient) for TEM mode (coaxial line) or TE 10 mode (rectangular waveguide) were measured using a vector network analyzer (Agilent Technology, 8722ES) by the full-two-port method. The values of µ r (µ r = µ r jµ r, j = 1) and ε r (ε r = ε r jε r) were calculated from the data of both S11 and S 21. The return loss R for each sample thickness was calculated from the complex reflection coefficient Γ using the relation R = 20 log 10 Γ. Figure 1: Frequency dependences of ε r and ε r for composites made of polystyrene resin and sendust particles of various sizes. 3. RESULTS AND DISCUSSION 3.1. Frequency Dependences of µ r and ε r for the Composites Made of Sendust Particles of Various Sizes Figure 1 shows the frequency dependences of ε r and ε r for the composites made of granular sendust particles of 20, 10, or 5 µm diameter. The volume mixture ratio of sendust was fixed at 53 vol%. The values of ε r and ε r decreased as the particle size decreased. This is because the sendust particles were more likely to be isolated from each other in the polystyrene resin as the particle size decreased. To investigate the dispersion of the sendust particles, the surfaces of the composites were observed using an optical microscope, as shown in Fig. 2. It is found from Figs. 2 and that sendust particles of 20 and 10 µm diameter are partly in contact with each other and form clusters. On the other hand, sendust particles of 5 µm diameter were dispersed uniformly and isolated from each other in the polystyrene resin, and the contact between sendust particles decreased as shown in 20 µ m sendust 10 µ m sendust (c) 5 µ m sendust Figure 2: Surface optical microphotographs of composites made of polystyrene resin and granular sendust particles. The sizes of the sendust particles are approximately 20 µm 10 µm and (c) 5 µm.

3 PIERS ONLINE, VOL. 4, NO. 8, Fig. 2(c). Therefore, the values of ε r were smaller, because the conductivity of the composite made of sendust particles of 5 µm diameter was lower. The values of ε r are determined by the capacitance between sendust particles. If the sendust particles are isolated, the distance between the particles becomes large and the capacitance between the particles decreases. Thus, the composite made of sendust particles of 5 µm diameter had low values of ε r due to the isolation of the particles. Figure 3 shows the frequency dependences of µ r and µ r for the composite made of sendust particles of various sizes. In the low-frequency range, the values of µ r were largest for the composite made of sendust particles of 10 µm diameter and decreased in the order of particles with 20 and 5 µm diameter. Meanwhile, the frequency at which µ r begins to decrease increased as the particle size of sendust decreased. With relation to the frequency dependence of µ r, the frequency at which µ r was maximum increased as the particle size of sendust became small and the value of µ r increased rapidly with frequency for the composite made of sendust particles of 5 µm diameter. It is considered that the isolation of the sendust particles is one factor that explains the reason why the frequency dependence of µ r is different for different sendust particle sizes [7, 8]. The scattering of an electromagnetic wave depend on the dispersion state of the sendust particles; thus, the frequency dependences of µ r are different. This phenomenon has been reported for a composite made of ferrite particles isolated in a SiO 2 medium [8]. For this composite, µ r had low values in the low-frequency range, was almost constant with increasing frequency then decreased rapidly at high frequencies. Moreover, the value of µ r increased rapidly and the frequency at which µ r was maximum increased as the ferrite particles became increasingly isolated. This frequency dependence of µ r is similar to that observed in the composite made of small sendust particles. Therefore, it is speculated that the difference in the frequency dependence of µ r for the composites made of sendust particles of various sizes is due to the isolation of the sendust particles in the medium of polystyrene resin. Figure 3: Frequency dependences of µ r and µ r for composites made of polystyrene resin and sendust particles of various sizes Frequency Dependences of µ r and ε r for the Composite Made of Granular Sendust, Sendust Flakes, or Granular Permalloy Figures 4 and 5 respectively show the frequency dependences of ε r and µ r for the composites made of granular sendust, sendust flakes, or granular permalloy. The particle size and volume mixture ratio of the granular sendust and permalloy are 10 µm and 53 vol% and those of sendust flakes are 20 µm and 57 vol%, respectively. The values of ε r and ε r for the composite made of sendust flakes were larger than those of the granular sendust and granular permalloy, as shown in Fig. 4. This result shows that flake type particles easily coalesce, because flake type particles are in contact at planes while granular type particles are in contact at points. Therefore, the values of ε r and ε r for the composite made of sendust flakes increased for the same reason as discussed in Section 3.1. The composite made of granular permalloy had higher values of µ r than that made of granular sendust at frequencies above 1 GHz. In addition, the frequency at which µ r for the composite made of granular permalloy has a maximum value was higher. However, the frequency dependence of µ r for the composite made of granular permalloy roughly agreed with that for the composite made of granular sendust. On the other hand, the value of µ r for the composite made of sendust flakes

4 PIERS ONLINE, VOL. 4, NO. 8, decreased rapidly compared with that for the composite made of granular sendust. In addition, µ r had a local maximum at 150 MHz and 3.5 GHz for the composite made of sendust flakes, as shown in Fig. 5. From the above results, it was found that the frequency dependence of µ r is different between sendust flakes and granular sendust in spite of the magnetic material being same. This is because the magnetic properties of sendust flakes, such as the crystallomagnetic anisotropy, are different from those of granular sendust, and the response of the electromagnetic wave to the composite varies with particle shape. Therefore, the different frequency dependence of µ r was observed for different particle shapes. Figure 4: Frequency dependences of ε r and ε r flakes and granular permalloy. for composites made of granular sendust, sendust Figure 5: Frequency dependences of µ r and µ r flakes and granular permalloy. for composites made of granular sendust, sendust 3.3. Absorption Characteristics of Composites Made of Sendust or Permalloy Figure 6 shows the absorption center frequency f 0 and the normalized 20 db bandwidth (the bandwidth corresponding to the return loss of less than 20 db divided by f 0 ) for the composites made of granular permalloy, granular sendust of 5 µm diameter, and that of 10 µm diameter. The value of 20 db corresponds to the absorption of 99% of the electromagnetic wave power. These three composites had a return loss of less than 20 db in the frequency range from 1 to 3 GHz. The sample thickness for which the return loss was less than 20 db was relatively thin and both composites had a bandwidth of more than 10%. Compared with the composite made of sendust particles of 10 µm diameter, f 0 for the composite made of sendust particles of 5 µm diameter was high and the sample thickness was thin. Although f 0 for the composite made of granular permalloy was lower than that for the composite made of granular sendust at a sample thickness 4 mm, the

5 PIERS ONLINE, VOL. 4, NO. 8, composite made of granular permalloy exhibited a return loss of less than 20 db at a sample thickness 3 mm, as shown in Fig. 6. Figure 6: Values of normalized 20 db bandwidth and center frequency f 0 for composites made of polystyrene resin and granular permalloy, granular sendust of 5 µm diameter, and granular sendust of 10 µm diameter. To investigate the absorption characteristics for these three composites, the measured values of µ r and the calculated values of µ r that satisfy the nonreflective condition given by Equation (1) are shown in Fig. 7 [9]. 1 = ( µ r/ε r tanh γ 0 d ) µ rε r (1) Here, γ 0 is the propagation constant in free space and d is the sample thickness. The value of ε r used for calculation is independent of frequency and the same as the measured value for each composite. ε r is assumed to be zero. For the sample thickness of 4 mm, the lines showing the values of µ r that satisfy the nonreflective condition moved to a higher frequency with decreasing value of ε r. Thus, f 0 for the composite made of sendust particles of 5 µm diameter with ε r = 16 was greater than that for the composite made of sendust particles of 10 µm diameter with ε r = 23. For the sample thickness of 3 mm, the calculated line for ε r = 16 and d = 3 mm intersected all the measured values of µ r in the frequency range from 3 to 4 GHz, as shown in Fig. 7. However, the calculated line only intersected the measured values of µ r for the composite made of granular permalloy, as shown in Fig. 7. Thus, the composite made of granular permalloy exhibited a return loss of less than 20 db for the sample thickness of 3 mm, because µ r for the composite made of granular permalloy had high values at frequencies above 1 GHz. Figure 7: Measured and calculated values of µ r and µ r. Plots show measured values for composites made of granular sendust and granular permalloy and lines show values calculated using Equation (1). The composites made of sendust flakes and granular sendust particles of 20 µm diameter did not exhibit a return loss of less than 20 db at frequencies above 1 GHz, although these two

6 PIERS ONLINE, VOL. 4, NO. 8, composites exhibited a return loss of less than 20 db at several hundred MHz. This is because these composites had high values of ε r and ε r and did not satisfy the nonreflective condition at frequencies above 1 GHz Frequency Dependence of µ r and Absorption Characteristics at Frequencies above 10 GHz Figure 8 shows the frequency dependences of µ r and µ r for the composites made of polystyrene resin and sendust particles of 5 µm diameter or permalloy particles of 10 µm diameter in the frequency range from 12.4 to 40 GHz. The values of µ r for both composites decreased and became less than unity at frequencies above 10 GHz. The value of µ r for the composite made of sendust particles of 5 µm diameter was minimum near 14 GHz and increased with increasing frequency up to 40 GHz. Although the frequency dependence of µ r for the composite made of permalloy particles of 10 µm diameter was similar to that of the composite made of sendust particles of 5 µm diameter, its minimum value of µ r was lower and the frequency of the minimum µ r value was higher. Figure 8: Frequency dependences of µ r and µ r for composites made of granular sendust particles of 5 µm diameter and granular permalloy of 10 µm diameter. To explain this frequency dependence of µ r, two reasons might be considered. The first reason is the natural magnetic resonance of soft magnetic materials. Soft magnetic materials have a resonance frequency in the frequency range from several hundred MHz to GHz, and the resonance frequency of a composite is higher than that of a pure magnetic material. The value of µ r decreases rapidly near the resonance frequency and becomes less than unity. Thus, the values of µ r for both composites are less than unity and increase as the frequency increases far from the resonance frequency. Another reason is the generation of magnetic moments by the eddy current flowing on the surface of soft magnetic material particles, because soft magnetic materials are conductive. The phenomenon that the values of µ r become less than unity has also been observed in composites made of polystyrene resin and conductive particles such as aluminum [10]. In this case, the value of µ r decreases upon increasing the volume mixture ratio of conductive particles and that of µ r decreases in inverse proportion to the particle size of the conductive particles. These results were in approximate agreement with Equations (2) and (3) obtained by a qualitative theoretical estimation. µ r = 1 V (2) µ r = 2V δ a (3) Here, V is the volume mixture ratio of the conductive particles, δ is the skin depth and a is the radius of the conductive particles. It is found from Equation (2) that the value of µ r is independent of the skin depth and frequency and only depends on the volume mixture ratio of the soft magnetic material. Applying Equation (2) to the composites made of sendust particles and permalloy particles gives µ r = 0.47 for V = On the other hand, the minimum value of µ r for the composite made of sendust particles of 5 µm diameter is approximately 0.51 and that for the composite made of permalloy particles of 10 µm diameter is approximately In addition, the value of µ r for the composite made of sendust particles of 5 µm diameter was largest and decreased

7 PIERS ONLINE, VOL. 4, NO. 8, in inverse proportion to the particle size at frequencies above approximately 6 GHz, as shown in Fig. 3. This particle-size dependence of µ r qualitatively agreed with Equation (3). These results suggest that the frequency dependence of µ r in the high frequency range might be explained by the generation of a magnetic moment. However, the reason for the frequency dependence of µ r in the high-frequency range is uncertain and further investigation is necessary. The composite made of sendust particles of 5 µm diameter exhibited a return loss of less than 20 db in the frequency range from 33 to 39 GHz for sample thicknesses of 2.5, 2.8 and 2.9 mm, although the bandwidth was narrow (approximately 1.4%). It was found that the composite made of sendust particles of 5 µm diameter can operate in the frequency range of not only several GHz but also above 30 GHz. Figure 9: Values of return loss for the composite made of sendust particles of 5 µm diameter. The volume mixture ratio of sendust is 53 vol%. 4. CONCLUSIONS The frequency dependences of µ r and ε r for composites made of polystyrene resin and granular sendust particles depended on the sendust particle size. Therefore, the absorption characteristics were different for sendust particles of different sizes, and the composite made of small sendust particles absorbed more than 99% of electromagnetic wave power in the frequency range from 1 to 3 GHz. The frequency dependences of µ r and ε r and the absorption characteristics for the composite made of granular sendust particles were similar to those of the composite made of granular permalloy particles but were different from that made of sendust flakes. The values of µ r for the composites made of granular sendust particles of 5 µm diameter and granular permalloy particles of 10 µm were minimum at frequencies above 10 GHz and increased with increasing frequency. The return loss of the composite made of granular sendust particles of 5 µm diameter was less than 20 db in the frequency range from 33 to 39 GHz for a suitable sample thickness. REFERENCES 1. Kasagi, T., T. Tsutaoka, and K. Hatakeyama, Particle size effect on the complex permeability for permalloy composite materials, IEEE Trans. Magn., Vol. 35, No. 5, , Kasagi, T., T. Tsutaoka, and K. Hatakeyama, Negative permeability spectra in permalloy granular composite materials, Appl. Phys. Lett., Vol. 88, 17502, Liu, J. R., M. Itoh, T. Horikawa, E. Taguchi, H. Mori, and K. Machida, Iron based carbon nanocomposites for electromagnetic wave absorber with wide bandwidth in GHz range, Appl. Phys. A, Vol. 82, , Lim, K. M., K. A. Lee, M. C. Kim, and C. G. Park, Complex permeability and electromagnetic wave absorption properties of amorphous alloy-epoxy composites, J. Non-Cryst. Solids., Vol. 351, 75 83, Matsumoto, M. and Y. Miyata, Thin electromagnetic wave absorber for quasi-microwave band containing aligned thin magnetic metal particles, IEEE Trans. Magn., Vol. 33, No. 6, , 1997.

8 PIERS ONLINE, VOL. 4, NO. 8, Olmedo, L., G. Chateau, C. Deleuze, and J. L. Forveille, Microwave characterization and modelization of magnetic granular materials, J. Appl. Phys., Vol. 73, No. 10, , Nakamura, T., T. Tsutaoka, and K. Hatakeyama, Frequency dispersion of permeability in ferrite composite materials, J. Magn. Magn. Mater., Vol. 138, , Kusunoki, H., T. Kondo, K. Hiraki, K. Takada, and S. Yoshikado, Grain-size dependence of the characteristics of electromagnetic wave absorbers of ferrite-sio 2 composites, Key Eng. Mater., Vol. 248, , Musal, Jr., H. M. and H. T. Hahn, Thin-layer electromagnetic absorber design, IEEE Trans. Magn., Vol. 25, No. 5, , Yamane, T., A. Nishikata, and Y. Shimizu, Resonance suppression of a spherical electromagnetic shielding enclosure by conductive dielectrics, IEEE Trans. on Electromagnetic Compatibility, Vol. 42, , 2000.